“National Champions and M&A deal duration”

(1)

Master Thesis

“National Champions and M&A deal duration”

Groningen, 27-09-2010

Author: Supervisor:

Mariëlle Booij Dr. P. Rao Sahib

S1453432 P.Rao.Sahib@rug.nl

(2)

1. Introduction

A number of recent EU merger cases have attracted a lot of media attention because of suspicion about protectionist behaviour of member states involved in these mergers (Motta and Ruta, 2007). European governments are accused of supporting large national firms in their merger and acquisition (M&A) activities. They tend to favour national mergers in order to keep these big players under domestic control and to bring them in a better position to compete with foreign firms in world markets (Haufler and Nielsen, 2008). The policy strategy of creating so-called “national champions”, promoting national rather than international mergers, has recently come to the forefront of the European policy debate (Calzolari and Scarpa, 2009). Proponents of national champion’s strategies argue that these policies enable firms to obtain a certain critical mass (economy of scale) in order to survive in today’s highly competitive market. Additionally these national champions often play a major role in the domestic economy. They are large buyers and suppliers of domestic goods and services. It is argued that growth and innovation in these sectors will spur growth and innovation in a wide range of complementary sectors. Therefore these sectors are considered strategically important and are subject to national champion policies (Geroski, 2005). Some EU countries, including France, Sweden and the UK, have even incorporated these policies in their national merger guidelines. According to these guidelines a merger can serve the national interest by increasing the market share of domestic firms in foreign markets (Haufler and Nielsen, 2008).

(4)

Dikova et al. (2010) investigated the effect of institutional differences and organizational learning on the length of the pre-acquisition phase. This study will contribute to this line of research by focusing on the highly debated industrial policy of promoting national champions, as this strategy could also have an important influence on the pre-acquisition process. We will focus on domestic M&A transactions entered into by a national champion. Assuming that European governments support these national champions in their M&A activities, it could be expected that “national champion M&A transactions” face a shorter deal duration than other comparable European M&A transactions.

As national governments seem to have a major influence in the pre-acquisition process, the characteristics of the national government itself could also determine the length of the pre-acquisition period. This study hypothesizes that the “power” of the home country of the acquiring firm will have a negative effect on the deal duration. It could be that more 'powerful' European countries will have a bigger influence and bigger say on the outcome and length of the acquisition process. As European governments support their national champions in their M&A activities, this study hypothesizes that the more powerful the national government, the shorter the deal duration.

At a firm-level another, possibly important determinant of deal duration is the acquisition experience of the acquiror. Frequent acquirers are more likely to have a successful M&A transaction because they have had the opportunity to learn from previous acquisitions (Al-Laham et al, 2010). Firms engaged in M&A activity develop acquisition process knowledge which can be used to develop practices and routines that facilitate the completion of future acquisitions (Collins et al. 2009). Vermeulen and Barkema (2001) argue that acquisition experience makes firms more flexible and adaptable to new and changing circumstances, which will increase the survival rate of future transactions and make early terminations less likely. Therefore this study hypothesizes that experienced firms have gained acquisition process knowledge which facilitates future acquisition processes and shortens the deal duration.

(5)

2. Theory and Hypotheses

2.1. The Acquisition Process

In 'How are firms sold?' Boone and Mulherin (2007) give an example of an acquisition process. Usually the process begins with a selling firm which hires an investment banker to determine the number of potential bidders to contact. These bidders will receive non-public information in exchange for a signed confidentiality agreement and are asked to indicate preliminary interest. Subsequently the interested bidders submit a binding sealed offer from which the winning firm will be selected. Only this firm will make a public offer for the selling firm. After that both firms will sign an initial merger agreement and enter the period of a public takeover (Hotchkiss et al., 2005) In this second period the firms continue to receive new deal information which can give the firms the incentive to renegotiate a part of the agreement or all of it. Therefore it usually takes several months to complete the public takeover process. This allows a lot of time for intervening events to complicate the completion of the acquisition. Especially for the acquiring firm a termination of the acquisition is very costly. It incurs substantial up-front costs in finding the appropriate partner and preparing the offer and additionally it foregoes the opportunity to acquire an alternative target (Bainbridge, 1990). This study focuses on the second period of the acquisition process: the period between the public announcement date and the resolution date (completion). Figure 1 visualizes the acquisition process.

Figure 1. Acquisition process. (Source: Boone and Mulherin, 2007) 2.2. National Champion strategies

(6)

global market (McGuire, 2006).

Proponents of national champion strategies argue that these policies enable firms to obtain a certain critical mass (economy of scale) in order to survive in today’s highly competitive market. As firms are increasingly subject to strong international competition, governments feel obligated to protect some critical national industries. These national champions often play a major role in the domestic economy. They are large buyers and suppliers of domestic goods and services. It is argued that growth and innovation in these sectors will spur growth and innovation in a wide range of complementary sectors. Therefore these sectors are considered to be strategically important and are subject to national champion policies (Geroski, 2005). Additionally, governments fear that if these national champions fall under foreign ownership, production will be moved away from the home country, which could destroy the relationships with local suppliers and jeopardize the position of domestic workers (Südekum, 2008). According to Südekum (2008, 2010) falling trade costs can also explain why governments are sensitive to national champion strategies. When a government is faced with the choice between a domestic and a foreign takeover, it will normally choose the option which maximizes national welfare. A foreign firm that acquires a domestic firm will improve its competitive position by the savings on trade costs. If these lower trade costs result in lower prices consumer welfare will increase and if prices remain constant producer welfare will increase due to higher profits. Either way the welfare effects will be positive. However, as Südekum (2008) argues, current trade costs are falling due to the integration process. Therefore the consumer and producer surplus and the welfare gains of the foreign takeover will decrease. As governments are biased against a foreign takeover, further integration will induce governments to promote national champions and to block a foreign takeover.

(7)

Sanofi-Synthelabo into the bidding process. Sanofi-Synthelabo ended up acquiring Aventis for $68 billion.

Apart from the European governments, the European Commission (EC) is also not free of suspicion of protectionism. European M&A legislation dates from the beginning of the 1990s with the objective to institute a system that ensures that competition in the common market will not be distorted (EC website). The EC recognizes that the completion of the internal market will continue to result in major corporate reorganizations, as low trade costs will make it profitable for firms to concentrate in one location (economy of scale). Although these reorganizations ought to be welcomed, the EC wants to ensure that this process does not result in lasting damage to competition (Council Regulation (EC) No 139/2004). The EC prevents mergers that significantly increase the concentration level in any given industry, as this will lead to anticompetitive practices at the expense of the consumers. As sound as this official position may seem, empirical research has been unable to find support for this position (Aktas et al., 2004). Several studies have (Atkas et al., 2004, Neven and Röller, 2002 and Duso et al., 2007) tested whether the nationality of the bidder determines the probability of regulatory intervention. In all three studies the effect turned out to be significant: mergers initiated by foreign bidders faced a higher probability of intervention. Evidence indicates that the EC attempts to shelter inefficient European firms from foreign competition. European M&A regulation procedures are explained in Box 1 (Aktas et al. 2004).

Box 1. European M&A regulation

Scope of intervention-Mergers and acquisitions fall under the jurisdiction of the European Union when the following three conditions are met: the total world-wide gross sales of all concerned firms exceed €5 billion; the European individual gross sales of at least two of the concerned firms exceed €250 million and companies concerned achieve more than two-thirds of their community-wide turnover within one and the same member state.

Juridical competence-The decisions of the European Commission (EC) are final and have the highest juridical authority.

(8)

According to the Competition Policy Review Panel a 'national champion' can be defined as a domestically-based company that has become a leading competitor in its global market. This study focuses on the M&A activity undertaken by a national champion. It is assumed that governments, when a national champion is involved, tend to favour a national merger instead of an foreign takeover. By creating a strong national player, governments try to survive in today’s highly competitive market. Earlier research on national champions has remained theoretical, in the sense that it has tried to explain why governments were trying to create national champions and what the market effects of this policy would be in terms of competition and innovation (e.g.Calzolari and Scarpa, 2009; Geroski, 2005; Haufler and Nielsen, 2008; McGuire, 2006; Motta and Ruta, 2007; Südekum, 2010). To the best of the author's knowledge this is the first empirical research that tries to quantify a national champion transaction. As governments seem to be in favour of national M&A when a national champion is involved, this study defines a “national champion transaction” as a transaction in which the target and the bidding firm are from the same European country. To exclude small transactions, which are not likely to be subject to national champion strategies, the database consists only of transactions which were under scrutiny by the European Commission (EC). Transactions need to be approved by the EC when: the total world-wide gross sales of the firms (world-wide gross sales of bidder plus target) is greater than €5 billion, the European individual gross sales of the two firms is greater than €250 million and when the two firms achieve more than two-thirds of their community-wide turnover within one and the same member state. Only these transactions fall under the EU authority with the objective to investigate whether these combinations could possibly distort the competition in the common market. As a consequence the database only consists of firms large enough to be subject to national champion strategies. Additionally the available market share data of the selected firms show that the database consists of industry-leading firms. As a result, by the definition of the Competition Policy Review Panel, each transaction in the database can be considered “national champion transaction”. To determine whether “national champion transactions” face a shorter deal duration than other comparable M&A transactions a second database was created with “comparable” transactions. In order to get meaningful results these transactions are exactly the same in terms of industry and firm characteristics. To prevent that location effects will influence the results, the comparison database only consists of transactions made in the Europe. The only and key difference with the “national champion transaction” database is that the comparison database exclusively selected transactions in which the target and the bidder firm are NOT from the same European country.

(9)

champion. Therefore this study hypothesizes that:

Hypothesis 1: European national champion M&A transactions face a shorter deal duration (period between deal announcement and deal completion) than other comparable European M&A transactions.

2.3. Country power

International trade liberalization has exposed European countries to fierce foreign competition. Since direct protection of domestic firms could possibly lead to retaliation by the World Trade Organization (WTO) and major costs in terms of economic inefficiency, governments nowadays turn to other adjustment policies in order to reposition firms. Creating national champions can be seen as one of these policies. Based on some recent EU merger cases, especially the five large EU countries seem to be very active in these kinds of policies. France, Germany, Italy, Spain and the United Kingdom have been more than once criticized for their protectionist attitude in some merger cases. This is not very surprising as these are exactly the countries facing the biggest problems concerning innovation and productivity (McGuire, 2006).

Additionally these countries have the greatest “power” to influence the acquisition process towards their preferred outcome. They have more “country power” in terms of population size, GDP and EU influence. Therefore this study hypothesizes that the “power” of the home country of the acquiring firm could also be an important determinant of the deal duration. As the suspicion exists that European governments support their national champions in their M&A activities, it could be argued that the more “powerful” the country the more effective this support will be, and therefore the shorter the deal duration. Powerful countries have the ability to influence the acquisition process strongly and may even have the power to determine the outcome.

Hypothesis 2: The “power” of the home country of the acquiror will have a negative effect on the deal duration.

2.4. Acquisition experience

(10)

regarding due diligence, deal negotiation and financing (Collins, 2009). They will develop routines for screening and purchasing the target. According to Hayward (2002) learning from experience is a “dynamic process in which firms a) engage in experiences, b) draw inferences from them and c) store the inferred material for future reference”. So when a firm accumulates acquisition experience it gains competence and expertise in their developed skills or routines (Haleblian et al., 2006) and will be more effective in applying these in future acquisitions. Therefore frequent acquirers are more likely to have a successful M&A transaction because they have had the opportunity to learn from previous acquisitions (Al-Laham et al, 2010). Acquisition experience will increase the survival rate of future acquisitions and make early terminations less likely. Therefore this study hypothesizes that experienced acquirers have developed skills and routines that help them to shorten the pre-acquisition period.

(11)

3. Methodology

3.1 Data

The “national champion transactions” were selected from the Database On Mergers in Europe (DOME) compiled at the Kiel Institute for World economics (1997-2000) and from the archive of the website of the European Commission (2000-2009). Both databases cover M&A transactions which were under supervision of the European Commission. The transactions in which the target and the bidder firm are from the same European country were selected. Subsequently, additional data on these selected transactions were obtained from Zephyr (a database which contains information about M&A transactions with links to company-specific financial information), Amadeus (European financial database) and EUROSTAT (statistics on the EU and candidate countries). The resulting database consists of 88 completed European national champions transactions in the period 1997-2010 with information of each transaction on: acquiror name, acquiror country code, target name, target country code, date announced, date completed, deal duration, deal type, deal status, acquiror turnover, acquiror NACE code, acquiror M&A experience, population size, one of the five largest European countries, GDP and seats in European Parliament. If the data was available also information about deal value, target turnover, target NACE code and market share was also included. Definitions of the different terms and variables are given in table 1 at the end of the Measures section.

(12)

comparable transactions that are quite similar with the national champion transactions in terms of industry (NACE-code) and firm characteristics (firm size or same acquiror). The key difference is that this comparison database only includes deals in which the target and acquiror firm are from different home countries.

3.2. Measures

The dependent variable of this study is the deal duration of the pre-acquisition period. This is calculated by the difference in days between the deal announcement date (the day the acquiring firm has made an public offer and has signed an initial agreement with the target) and the deal completion date (at this day the acquisition is completed and the target formally belongs to the acquiror).

The first independent variable is the dummy variable national champion transaction. This takes the value of one when the transaction concerns a “national champion transaction” (acquiror and target firm are from the same home country and are under EC scrutiny) and zero when the transaction concerns a comparable transaction (acquiror and target firm are from different home countries). The second independent variable, country power, is measured by several indicators. We measured “country power” by measuring the size of the country, in terms of population size and by measuring the 'richness' of the country, in terms of GDP per capita. We expect that bigger and wealthier countries have more influence on the acquisition process. Additionally, as we mentioned before, especially the five large EU countries seem to be attracted to national champion strategies, therefore we also investigated whether the home country of the acquiror will effect the deal duration. Finally, as this study focuses on European transactions it can be expected that the greater the influence a country has in the European arena the greater the influence it will have on the acquisition process. The several indicators are measured as follows. The first indicator, population size, is measured by the number of residents of that particular country on the first of January of the year of the announcement of the M&A transaction. The second indicator determines whether the home country of acquiror firm is one of the five largest European countries. This is a dummy variable, Big Five, which takes the value 1 when the country concerned is the United Kingdom, Germany, France, Spain or Italy and zero otherwise. The third indicator is the Gross Domestic Product per capita at market prices, GDP, measured at the end of the year of the announcement of the M&A transaction. The last indicator of country power is EU influence, which is measured by the number of seats in the European Parliament (EP seats).

In the third hypothesis we test whether the acquisition experience of the acquiring firm will influence deal duration. Acquisition experience is measured by the number of completed M&A transactions conducted by the same acquiring firm prior to the focal transaction.

(13)

(14)

Table 1. Definitions of terms.

Term Definition

Acquiror name The name of the firm that is the acquiror in the corresponding transaction.

Acquiror country code The country code of the home country of the firm that is the acquiror in the corresponding transaction.

Target name The name of the firm that is the target in the corresponding transaction.

Target country code The country code of the home country of the firm that is the target in the corresponding transaction.

Date announced At this date the acquiror has made a formal offer to the target, or the companies involved in the deal confirm that the deal is to go ahead (Zephyr).

Date completed At this date the transaction is completed and the target legally belongs to the acquiror.

Deal duration Difference in days between the deal announcement date and the deal completion date.

Deal type Indicates which type of deal was made: acquisition or merger.

Deal status Indicates the stage of development of the deal (in this study only completed deals were included.)

Acquiror turnover The operating revenue of the acquiring firm.

Acquiror NACE code The four-digit NACE industry code of the acquiring firm. Nomenclature statistique des Activités économiques dans la Communauté Européenne; four-digit industry classification system designed by the European Union (Zephyr).

Acquiror M&A experience

Number of completed M&A deals conducted by the same acquiring firm prior to the focal transaction.

(15)

One of the five largest European countries.

When the home country of the acquiring firm is the United Kingdom, Germany, France, Spain or Italy

GDP per capita Gross Domestic Product per capita measured at the end of the focal year.

Seats in European Parliament

Number of seats in the European Parliament of the home country of the acquiring firm.

Deal value The consideration paid for actual stake acquired (Zephyr).

Target turnover The operating revenue of the target firm.

Target NACE code The four-digit NACE industry code of the target firm. Nomenclature statistique des Activités économiques dans la Communauté

Européenne; four-digit industry classification system designed by the European Union (Zephyr).

Market share Sales in Euros of the acquiror firm divided by the sales in Euros of top ten percent firms (in terms of sales) in the same four-digit industry times 100 percent.

Method of payment The payment form of the acquisition;

-by cash; refers to payment by cheque or transfer of funds. -by shares; the acquiror gives its own shares to the target.

Percentage acquired stake

The ownership stake sought by the acquiror in the target.

Acquiror listed Indicates whether the acquiring firm is publicly owned.

Target listed Indicates whether the target firm is publicly owned.

Industry dummy Dummy variable which controls for industry effects.

3.3. Methods

(16)

Model 1:

ε

β

+ + + + + + = ) ( 5 ) ( 4 ) ( 3 ) (% 2 ) ( 1 0 mmy Industrydu ed Targetlist sted Acquirorli ake acquiredst yment Methodofpa on Dealdurati

Model 2 estimates the effect of the independent variables National champion transaction, M&A experience and Country power on the dependent variable Deal duration. In this model Country power is measured by GDP per capita and EP seats. The two other indicators of Country power, Population size and Big Five, were excluded from the Model, due to very high correlation values and insignificant coefficients (will be explained later on in this section). Again, both databases have been used for this model. Not all European countries included in this database are members of the European Union, therefore there are a few missing values for the independent variable EP seats. This resulted in a total number of observations of 729. Model 2:

ε

β

+ + + + + = ) ( 4 ) ( 3 ) exp & ( 2 ) ( 1 0 EPseats ta GDPpercapi erience A M saction ampiontran Nationalch on Dealdurati

Model 3 estimates the effect of all the independent (except Population size and Big Five) and control variables on the dependent variable Deal duration. Both databases were included in this model. Again, due to several missing values of the control variable Method of payment the number of observations for this model is 188.

Model 3:

ε

β

+ + + + + + + + + + = ) ( 9 ) arg ( 8 ) ( 7 ) (% 6 ) ( 5 ) ( 4 ) ( 3 ) exp & ( 2 ) ( 1 0 mmy Industrydu etlisted T sted Acquirorli ake acquiredst yment Methodofpa EPseats ta GDPpercapi erience A M saction ampiontran Nationalch on Dealdurati

(17)

Model 4:

ε

β

+ + + + + + + + + = ) ( 8 ) arg ( 7 ) ( 6 ) (% 5 ) ( 4 ) ( 3 ) exp & ( 2 ) ( 1 0 mmy Industrydu etlisted T sted Acquirorli ake Acquiredst EPseats ta GDPpercapi erience A M saction ampiontran Nationalch on Dealdurati

Models 5 and 6 were included to check whether the two different selection methods for the comparison transactions will result in significantly different coefficients and thus have influenced the results. As mentioned in the Data section, the first method of selecting the transactions of the comparison database, was by selecting the transactions (in which the acquiror and target firm are NOT from same home countries) with the same acquiror as the corresponding acquiror of the national champion transaction and within the same two-digit industry code. This implies that the complete database with both the national champion and comparison transactions, include transactions with the same acquiror. To correct for unobserved firm effects, instead of using a panel regression, we included Model 5 and 6 to check whether the two selection methods gave significant different results. Therefore Model 5 estimates the effect of all independent and control variables on the Deal duration of only national champion and comparison transactions based on the same acquiror selection method. To compare these results with the other selection method, Model 6 estimates the effect of all independent and control variables on the Deal duration of only national champion and comparison transactions based on the firm size method. On the basis of these models it will be checked whether the two selection methods influenced the results. To measure these models, two new dummy variables were created; Same acquiror that takes the value one when the focal transaction is a national champion transaction or a comparable transaction based on the first selection method and zero if the focal transaction is a comparable transaction based on the second selection method and Different acquiror that takes the value one when the focal transaction is a national champion transaction or a comparable transaction based on the second selection method and zero if the focal transaction is a comparable transaction based on the first selection method. The number of observations for Model 5 and 6 are respectively, 82 and 138. Due to the fact that the variables EP seats and Method of payment have some missing values, the number of transactions of both databases (national champion transaction and comparison database) and the number of observations of models 5 and 6 decreased.

Model 5:

Deal duration (if same acquiror = 1) =

(18)

Model 6:

Deal duration (if different acquiror =1) =

ε

β

+ + + + + + + + + + ) ( 9 ) arg ( 8 ) ( 7 ) (% 6 ) ( 5 ) ( 4 ) ( 3 ) exp & ( 2 ) ( 1 0 mmy Industrydu etlisted T sted Acquirorli ake acquiredst yment Methodofpa EPseats ta GDPpercapi erience A M saction ampiontran nationalch

Due to the cross-sectional nature of the data, all the models mentioned above are being tested by using the Ordinary Least Squares (OLS) regression technique. When using this regression technique we need to make certain assumptions:

1. The expected value of the dependent variable(y) depends on the values of the dependent variables(x) and the unknown parameters (

β

). Research model should have following form:

e xk k x

y=

β

0+

β

1( 1)+....+

β

( )+ with E(e)=0.

2. For each observation the random variable y has the same variance. “The variance of the probability density function of y does not change with each observation” (Hill et al., 2008). 3. Values of the error term e should all be uncorrelated and have zero variance.

4. “The values of each independent variable, x, are random and are not exact linear functions of the other explanatory variables” (Hill et al., 2008).

5. The values of the error term, e, should be normally distributed about their mean (Hill et al., 2008).

(19)

a set of explanatory variables. The null-hypothesis of this test assumes that the error variance is constant and heteroskedasticity is not present. The downside of using the Breush-Pagan test is that this test presupposes that we have knowledge of what variables will appear in the error variance function and which are the relevant explanatory variables (Hill et al., 2008). In this case we do not exactly know for which explanatory variables heteroskedasticity exists. Therefore we used an additional test, the White test. Both the Breush-Pagan and the White test were performed on all independent and control variables used in Model 3 (results can be found in Appendix 1). The results show that there is not enough evidence to reject the null hypothesis (constant variance), so we can safely say that the data used in this study does not suffer from heteroskedasticity.

The third assumption of the OLS regression technique indicates that the covariance between two random errors is zero. The size of the error for one observation should not affect the size of the error for another observation (autocorrelation). Although autocorrelation is unlikely to occur in cross-sectional datasets, it is necessary to check this assumption for the results to be valid. In order to test for autocorrelation we computed a residual correlogram for the complete deal duration equation, Model 3 (See Appendix 2). A residual correlogram plots series of correlations against the time interval. Each dot represents the correlation between residuals that are one period apart, two periods apart, three periods apart etc. The dark area is the 95% confidence band (Hill et al., 2008). All dots lie inside the boundary, meaning that the correlations between the errors are not significantly different than zero; hence the data does not suffer from autocorrelation.

(20)

Table 2. Correlations between independent, dependent and control variables. Variables 1 2 3 4 5 6 7 8 9 10 11 12 1. National champion Transaction 1.00 2. M&A experience -0.08 1.00 3.Population size 0.15 0.04 1.00 4. Big five 0.15 0.06 0.93 1.00 5. GDP -0.09 0.03 -0.30 -0.3 1.00 6. EP seats 0.l5 0.04 0.99 0.92 -0.3 1.00 7. Method of payment -0.03 0.06 -0.02 -0.01 0.06 -0.02 1.00 8. % Acquired stake 0.08 -0.14 -0.17 -0.15 0.04 -0.17 0.02 1.00 9. Acquiror listed -0.17 0.13 -0.11 -0.05 0.12 -0.11 -0.06 0.04 1.00 10. Target listed 0.14 -0.02 0.08 0.08 0.01 0.08 -0.06 -0.49 -0.09 1.00 11. Industry dummy -0.01 0.07 -0.03 -0.06 0.03 -0.04 0.08 0.01 -0.10 0.03 1.00 12. Deal duration 0.23 -0.08 0.05 0.05 0.06 0.05 -0.05 0.02 -0.05 0.15 0.01 1.00

Table 3. Descriptive statistics for independent, dependent and control variables.

Variables ₁ ₂ ₃ ₄ ₅ ₆ ₇ ₈ ₉ ₁₀ ₁₁ ₁₂ Number of observations 756 756 756 756 741 742 218 698 756 737 756 756 Mean _0.12 _27.81 _42900000 _0.59 _27460.86 _54.91 _0.89 _71.17 _0.68 _0.19 _0.55 _46.92 Standard deviation 0.32 74.24 29100000 0.49 7115.06 31.58 0.32 36.58 0.47 0.39 0.50 155.50 Minimum ₀ ₁ ₃₀₇₆₇₂ ₀ ₄₉₀₀ ₆ ₀ _0.02 ₀ ₀ ₀ ₀ Maximum 1 955 82500000 1 80500 99 1 100 1 1 1 2307

(21)

variables are no true determinants of deal duration; therefore the models do not suffer from the omitted-variable bias.

For the last assumption to hold, the regression errors should be normally distributed. Hypothesis tests and interval estimates rely on this assumption. As we cannot observe the true random errors, we checked the normality of the least squares residuals (Hill et al. 2008). Histograms of the regression residuals and the Jarque-Bera tests indicated that all models suffered from nonnormality of the regression error. To overcome this problem of nonnormality of the error distribution, improving the model by using another functional form can be considered. A functional form that is applicable for this study is the log-linear model, which implies that the dependent variable was transformed by the natural logarithm. As our dependent variable several times takes the value zero, the transformation resulted in a lot of missing values. Still, we decided to use the log-linear form for model 1-4, to check whether this would solve the problem of nonnormality. Although the histograms of the regression errors seem somewhat more symmetrical, the null-hypothesis of normality of the error distribution is rejected for all four models. Indicating that by using the log-linear model the problem of nonnormality is not solved. However, theory indicates that the least squares estimators of the linear OLS regression are approximately normally distributed in large samples. In this study the samples used to measures the different linear models can be considered large enough, based on the rule of thumb (number of observations minus the number of unknown parameters is larger than 50). Therefore we chose to use the linear OLS regression technique.

Finally, all variables were checked for outliers. The dependent variable Deal duration and the independent variable M&A experience show some remarkable high observations (see box plots Appendix 3). We checked whether these outliers influenced the results, by removing them from the data. As the regression results for all models remained the same, we chose to keep the outliers in the sample.

(22)

4. Results

Table 4. Linear OLS regression results Model 1-6.

Dependent Variable: M&A Deal Duration

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Intercept _(25.98)39.57 67.88** _(28.91) _(40.21)4.99 89.96*** _(35.53) _(109.33)16.57 _(47.22)7.22 National champion transaction 47.16** (18.59) 41.97*** (15.95) 54.40*** 19.49 51.71** (22.65) 41.50** (18.24) M&A experience _(-0.11)-0.02 _(0.08)-0.06 -0.02 _0.14 _(0.44)0.11 _(0.09)-0.05 GDP _(0.00)0.00 _(0.00)0.00 _(0.00)0.00 _(0.00)0.00 _(0.00)0.00 EP seats _(0.19)0.16 _(0.21)0.13 _(0.20)0.00 _(0.39)0.31 _(0.27)0.13 Method of payment -7.04 (17.81) -9.15 (18.66) -36.20 (33.46) -5.20 (22.86) % acquired stake 0.27 (0.17) 0.16 (0.18) -0.30 (0.19) 0.46 (0.37) 0.14 (0.22) Acquiror listed _(12.76)-6.29 _(13.33)-1.22 _(12.87)8.60 _(27.09)13.50 _(15.79)2.40 Target listed 33.59** _(13.99) _(14.40)25.15* _(17.90)13.27 _(24.67)4.73 _(17.43)28.01 Industry dummy _(11.44)-1.60 _(11.65)1.50 _(11.73)5.75 _(20.85)-11.99 _(15.26)-1.30 No. of observations 195 729 188 661 82 138 R-squared 0.0338 0.0148 0.0826 0.0269 0.1263 0.0864 Adjusted R-squared 0.0082 0.0093 0.0362 0.0150 0.0171 0.0222 N-K 189 724 178 652 72 128

*p<0.10, **p<0.05 and ***p<0.01 (standard errors in parentheses)

(23)

difficulty with using R-squared as a goodness-of-fit measure is that this value can be made large by adding more explanatory variables. To check whether the R-squared values in Models 3, 5 and 6 are high due to the fact that these models include the most explanatory variables, we also reported the adjusted R-squared values for each model. For all models the adjusted R-squared values are rather low. It seems that Model 3 “fits” the data the best (Goodness-of-fit), this model has the highest adjusted R-square value. The low explanatory power of the six models can be seen as troublesome, however, as mentioned in the introduction this study contributes to a relatively new research field. To date there is very little literature on the pre-acquisition phase, therefore it is important in this early stage to look for possible determinants of deal duration and not so much to try to explain as much as possible variation in deal duration.

The results show that only one control variable, namely Target listed, produces a significant estimate. From model 1 and 3 we observe that transactions involving a publicly owned target, as expected, take longer to complete, as they are more complex. However this control variable loses its significance in the models 4, 5 and 6, when more explanatory variables are added. All other control variables do not have a significant influence on the dependent variable M&A deal duration. The insignificant results found for the control variable, Industry dummy, indicate that industry effects did not significantly influenced the results.

The results show no support for Hypothesis 2 and 3. Hypothesis 2 predicted that the “power” of the home country of the acquiror will have a negative effect on the dependent variable M&A deal duration. None of the reported coefficients in the models are significant. An F-test was performed to check whether the two country power variables, GDP and EP seats, in Model 3 jointly have any effect on deal duration. The p-value we found for this test was 0.47, indicating that we cannot reject the null hypothesis, which means that it is likely that country power has no effect on deal duration.

The same is applicable for hypothesis 3. Hypothesis 3 predicted that acquisition experience of the acquiror will have a negative effect on the deal duration. Although the four models show a very small negative relationship between Acquisition experience and M&A deal duration, all coefficients are not significant. So based on these results both hypothesis 2 and 3 are rejected. There seems to be no significant relationship between Country power and the M&A deal duration and no significant relationship between Acquisition experience and M&A deal duration.

(24)

and the dependent variable M&A deal duration for all models. However the signs of the coefficients are exactly the opposite of what we expected. From Model 3 (that includes all the explanatory variables and has the highest adjusted R-squared value) we can see that on average, the deal duration of a national champion transaction takes 41.97 days longer than the deal duration of other comparable European M&A transactions. This means that hypothesis 1 is rejected. Based on these results we conclude that, contrary to what we have expected, European national champion M&A transactions face a longer deal duration (period between deal announcement and deal completion) than other comparable European M&A transactions.

In Model 4 we excluded the control variable, Method of payment, from the regression equation of Model 3, as this variable strongly influenced the number of observations. The results show that by excluding this variable, the conclusions we draw for the hypotheses remain the same. Based on this model, again all hypotheses are rejected. Although there is a statistically significant relationship between the independent variable National champion transaction and the dependent variable M&A deal duration, the sign of the coefficient is exactly the opposite of what we expected. What is more interesting is that this is the only model with a significant intercept at a 1% level. The intercept can be interpreted as the average deal duration of a comparable transaction (National champion transaction=0) when all other explanatory variables are zero, which is 89.96 days.

(25)

5. Discussion

This study makes two contributions to existing literature. First it investigates a part of the acquisition process that has been ignored in literature. Earlier research has focused on the post-acquisition financial performance. However, analyzing the pre-acquisition period, the period between deal announcement and deal completion, can be of major importance to science and to managers. As firms make substantial up-front costs in preparing the definite offer and closing the deal (Bainbridge, 1990), gaining insight in what influences the length of this process can help managers to avoid prolonged deal-making.

Secondly, this research links M&A deal duration with national champion strategies. The strategy of creating national champions, by promoting national mergers instead of international takeovers, has recently come to the forefront of the European policy debate. However research on this subject has remained theoretical, in the sense that it tried to explain why governments try to create national champions and what the market effects of this policy will be in terms of competition and innovation (e.g.Calzolari and Scarpa, 2009; Geroski, 2005; Haufler and Nielsen, 2008; McGuire, 2006; Motta and Ruta, 2007; Südekum, 2010). This research is one of the first, or maybe even the first one, that tries to quantify a national champion transaction.

(26)

completed). Probably, M&A transactions will only be reported to the EC, when the negotiations between the firms involved were successfully finished. If the abandonment rate of national champion transactions (under scrutiny by the EC) is substantially lower than the abandonment rate of other comparable European M&A transactions, we can conclude that although national champion transactions have a relatively long deal duration (due to extensive investigation by EC), they are most of the time successfully completed. Unfortunately, this study could not find data for the abandonment rate of both databases. Future research could investigate this proposition or could overcome this problem by finding another way to define a national champion transaction.

This study also investigated whether M&A transactions with a more powerful home country had a shorter deal duration than transactions with a less powerful home country. We could not find any significant results. The last hypothesis tested whether acquisition experience will have a negative effect on deal duration. We expected more experienced acquirers to have developed skills and routines that help them to shorten the pre-acquisition period. Although some of the coefficients showed the expected direction, none of them were significant. These results lead to the conclusion that neither country power nor M&A experience are significant determinants of M&A deal duration.

(27)

Appendix 1. Heteroskedasticity

Table 1 appendix 1: heteroskedasticity tests.

(28)

(29)

Appendix 3. Boxplots

Boxplot Deal Duration

0 500 1,000 1,500 2,000 2,500

Deal duration

Boxplot M&A Experience

0 200 400 600 800 1,000

(30)

References

Abravanel R., and Ernst, D. 1992. Alliance and acquisition strategies for European national champions. McKinsey Quarterly, 2, 44-62.

Al-Laham, A., Schweizer, L. and Amburgey, T.L. 2010. Dating before marriage? Analyzing the influence of pre-acquisition experience and target familiarity on acquisition succes in the “M&A as R&D” type of acquisition. Scandinavian Journal of Management, 26, 25-37.

Aktas, N., de Bodt, E. and Roll, R. 2004. Market response to European regulation of business combinations. The Journal of Financial and Quantitative Analysis, 39 (4), 731-757.

Aktas, N., de Bodt, E. and Roll, R. 2007. Is European M&A regulation protectionist? The Economic Journal, 117, 1096-1121.

Anwar S.T. 2008. Creating a national champion or a global pharmaceutical company: a tale of French connection. Journal of Business and Industrial Marketing, 23(8), 586-596.

Aubert, C., Falck, O. and Heblich, S. 2008. Subsidizing national champions: an evolutionary perspective. CESifo working paper, No. 2380.

Bainbridge, S.M. 1990. Exclusive merger agreements and lock-ups in negotiated corporate acquisitions. Minnesote Law Review, 75(1), 239-334.

Boone, A.L. and Mulherin, J.H. 2007. How are firms sold? The Journal of Finance, 62(2), 847-875.

Calzolari, G., and Scarpa, C. 2009. Footlose monopolies: regulating a “national champion”. Journal of Economics & Management Strategies, 18(4), 1179-1214.

Cartwright, S. and Schoenberg, R. 2006. Thirty Years of Mergers and Acquisitions Research: Recent Advances and Future Opportunities. British Journal of Management, 17, S1-S5.

Collins, J.D., Holcomb, T.R., Certo, S.T., Hitt, M.A. and Lester, R.H. 2009. Learning by doing: Cross-border mergers and acquisitions. Journal of Business Research, 62, 1329-1334.

(31)

Evidence from the International Business Service Industry in 1981-2001. Academy of Management Best Conference Paper

Dikova, D., Sahib, P.R. and van Witteloostuijn, A. 2010. Cross-border acquisition abandonment and completion: The effect of institutional differences and organizational learning in the international business service industry, 1981–2001. Journal of International Business Studies, 41, 223-245.

Duso, T., Neven, D.J. and Röller, L.H. 2007. The Political Economy of European Merger Control: Evidence using Stock Market Data. Journal of Law and Economics, 50, 455-489.

Geroski, P. A. 2005. Competition policy and national champions.

www.competition-commission.org.uk/.../geroski_wifo_vienna_080305.pdf.

Haleblian, J., Kim, J.Y. and Rajagopalan, N. 2006. The influence of acquisition experience and performance on acquisition behavior: evidence from the U.S. commercial banking industry. Academy of Management Journal, 49(2), 357-370.

Haufler, A. and Nielsen, S.B., 2008. Merger policy to promote 'global players'? A simple model. Oxford Economic Papers, 60, 517-545.

Hayward, M.L.A. 2002. When do firms learn from their acquisition experience? Evidence from 1990-1995. Strategic Management Journal, 23(1), 21-39.

Hill, R.C., Griffiths, W.E. and Lim, G.C.2008. Principles of econometrics. John Wiley & Sons, Inc, Hoboken, US.

Holl, P. and Kyriazis, D. 1996. The determinants of outcome in UK takeover bids. International Journal of Economics and Business, 3(2), 165-184.

Hotchkiss, E., Qian, J. and Song, W. 2005. Holdups, renegotiation and deal protection in mergers. Working Paper , SSRN, http://ssrn.com/abstract=705365.

(32)

McDonald, M.L., Westphal, J.D. and Graebner, M.E. 2008. What do they know? The effects of outside director acquisition experience on firm acquisition performance. Strategic Management Journal, 29, 1155-1177.

McGuire, S. 2006. No more euro-champions? The interaction of EU industrial and trade policies. Journal of European Public Policy, 13, 887-905.

Moschieri, C. and Campa, J.M., 2009. The European M&A Industry: A Market in the Process of Construction. Academy of Management Perspectives, November, 71-87.

Motta, M. and Ruta, M. 2007. Mergers and National Champions. CESifo

Motta, M. and Ruta, M. 2008. A political economy model of merger policy in international markets. CEPR Discussion Paper, No. 6894.

Neven, D. and Röller, L.H. 2002. Discrepancies Between Markets and Regulators: an Analysis of the First Ten Years of EU Merger Control. HEI Working Paper No: 10/2002

Reus, T.H. and Lamont, B.T. 2009. The double-edged sword of cultural distance in international acquisitions. Journal of International Business Studies, 40, 1298-2506.

Südekum, J. 2008. National Champion versus Foreign Takeover. Ruhr Economic Papers, No. 66.

Südekum, J. 2010. National champions and globalization. Canadian Journal of Economics, 43(1), 204-231.

Swaminathan, V., Murshed, F. and Hulland, J. 2008. Value Creation Following Merger and Acquisition Announcements: The Role of Strategic Emphasis Alignment. American Marketing Association, 14, 33-47.

Vermeulen F. and Barkema H. 2001. Learning through acquisitions. The Academy of Management Journal, 44(3), 457-476.

(33)

Yasar, M. and Rejesus, R.M. 2005. Exporting status and firm performance: Evidence from a matched sample. Economic Letters, 88, 397-402.

“National Champions and M&A deal duration”

Master Thesis